Age-Associated Myeloid Biased Hematopoiesis Depends on Relative Decrease of Short-Term Hematopoietic Stem Cell

Aging in the human hematopoietic system is known to be associated with reduced self-renewal capacity, myeloid biased hematopoiesis, and increased incidence of hematological disorders. Amongst these changes, myeloid-biased hematopoiesis leads to decreased adaptive immune system function and increased propensity for myeloid malignancies in aged individuals.

Recent reports have supported the concept that myeloid-biased alteration is most likely due to cell intrinsic alterations in the hematopoietic stem cell (HSC) compartment and a clonal shift toward myeloid-biased HSCs. Recently, we demonstrated that Hoxb5 specifically marks HSCs with long-term self-renewal capacity (LT-HSC) within the bone marrow of young mice. Additionally, we further demonstrated that around 80% of the cells included within the immunophenotypic HSC fraction (defined by the cell surface markers: Lineage-c-Kit+Sca-1+Flk2-CD150+CD34-/lo) (hereafter referred to as pHSCs) are not LT-HSCs (Chen JY et al., Nature 2016). In this study, using our LT-HSC-specific reporter mouse model, we sought to understand the mechanisms underlying the myeloid-biased alteration with age.

First, to verify that Hoxb5 can be used as a LT-HSC-specific marker in aged mice, we utilized a transplantation assay in which 10 cells of Hoxb5⁺ or Hoxb5^- pHSCs isolated from 2-year-aged mice were transplanted with supporting bone marrow cells into lethally irradiated mice. Any mice showing long-term (≧16week) granulocyte reconstitution (≧1% generation in peripheral blood) were considered as a positive for long-term hematopoiesis. Refer to this criteria, continuous hematopoiesis was observed only within the Hoxb5⁺pHSC recipients, indicating that Hoxb5 marks LT-HSCs within the pHSC compartment throughout the mouse lifespan. Interestingly, the lineage output derived from transplanted aged-Hoxb5⁺-HSCs was not skewed towards the myeloid lineage when compared to their young counterparts (the frequency of myeloid lineage contribution in donor cells: Aged vs Young = 42.6% vs 50.5%, n.s.). Our findings suggest that the balance between myeloid and lymphoid lineage output of the LT-HSC (Hoxb5⁺ pHSC) persists throughout life.

To further confirm this hypothesis, we co-transplanted 10 cells of 2-year-aged Hoxb5⁺ pHSC with 10 cells of EGFP-overexpressed young Hoxb5⁺ pHSC along with supporting bone marrow cells into the same recipients. At 12 weeks posttransplant, peripheral blood (PB) analysis demonstrated that the frequency of myeloid output in both aged or young donor cells was nearly equivalent (Aged vs Young = 48.0% vs 43.2%, n.s.).

According to previous reports, myeloid-related genes are known to be enriched in HSC compartment with age. However, given our previous findings demonstrating that ~80% of the pHSC compartment consists of non-LT-HSCs, we conducted RNA-seq analysis for bulk pHSCs (a mixture of Hoxb5⁺ and Hoxb5^- pHSCs) and pure Hoxb5⁺ pHSCs isolated from young and 2-year-old mice respectively. Gene Set Enrichment Analysis showed that the previously defined geneset: preGM (granulocyte/macrophage progenitors)-specific genes (Pronk, C. J. et al., Cell Stem Cell 2007) are more highly expressed in aged pHSCs than in young pHSCs (NES = 1.25, q=0.067), whereas no significant gene enrichment of preGM-specific genes in aged Hoxb5⁺ pHSCs was observed compared to young Hoxb5⁺pHSCs (NES=-1.02, q=0.333).

Given that the frequency of the Hoxb5^- pHSC (which possesses only transient self-renewal and give rise to mainly lymphoid cell) in the pHSC compartment is decreased in aged mice (Aged vs Young = 43.3% vs 73.6%, p<0.0001), we hypothesized that the myeloid biased phenotypic change in the HSC compartment with age depends on a relative decrease of Hoxb5^- pHSCs. To verify this, we transplanted young Hoxb5⁺ pHSCs and Hoxb5^- pHSCs in a 5:5 ratio (the ratio between Hoxb5⁺ and Hoxb5^- pHSCs in aged mice) or 2:8 ratio (the ratio in young mice). PB analysis at 16 weeks post-transplant demonstrated that recipient mice transplanted with Hoxb5⁺ pHSCs and Hoxb5^- pHSCs in the 5:5 ratio have a greater myeloid lineage output than recipients of the 2:8 ratio (the frequency of myeloid lineage contribution in donor cells: 5:5 ratio vs 2:8 ratio = 31% vs 7%, p<0.05).

In conclusion, the present study demonstrates that myeloid biased hematopoiesis in the aged setting results from the alteration of the ratio between Hoxb5⁺ and Hoxb5^- pHSCs rather than cell-intrinsic change.